Design And Implementation Of User Behavior Analysis System Based On Communication Log

As network traffic explodes,more and more research is based on network flow analysis.The analysis of network flow is of great significance for improving network service quality and improving network management.This paper designs and implements a user behavior analysis system based on the communication log to complete the network flow analysis of the communication log format.The system implements three functions:network address calibration,user behavior analysis,server status monitoring.Specifically,the system includes the following sections:1.Design and implementation of network address calibration module.Based on the difference of communication information between the client and the server in the communication log,this paper designs and implements the network address calibration module.The module extracts the network log of the communication log,and performs the calibration of the client network address and the server network address based on the port,the uplink and downlink bytes,the number of sessions and the active information.The system uses Hadoop to build a distributed environment,MapReduce as a distributed computing engine,HDFS and HBase as the underlying data storage framework to complete the aggregation,statistics and calibration of the communication log and to realize the function of network address calibration.2.Design and implement user behavior analysis module.In this paper,considering the large number of users in the backbone network,this paper proposes a grouping algorithm based on active time distribution,dividing the center of population into groups,calculating the Euclidean distance between the user's active time distribution vector and the center point vector,So as to discover user groups with different active time distributions.In addition,this paper analyzes the usage preference and time of client users,calculates the service usage pattern of users in different periods,and divides groups according to the service usage pattern of users,so as to provide decision-making for different groups of users for directional marketing.3.Design and implement server status monitoring module.By observing the communication log and extracting the status indicators that can describe the status of the server,this paper uses the index fusion algorithm based on time window to calculate the service status of each time window of the server and find out the time window of abnormal service status to detect whether the server Normal service.System design and implementation of each module to complete the processing and analysis of the communication log.Realize the network address of the function of communication log calibration function;The user based on the active time distribution and business use of the grouping of the analysis of the partition;The server based on the time window of the state monitoring and analysis.Finally,the various functions of the experimental test and analysis.